Progressive label fusion framework for multi-atlas segmentation by dictionary evolution

Yantao Song, Guorong Wu, Quansen Sun, Khosro Bahrami, Chunming Li, Dinggang Shen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

Accurate segmentation of anatomical structures in medical images is very important in neuroscience studies. Recently, multi-atlas patch-based label fusion methods have achieved many successes, which generally represent each target patch from an atlas patch dictionary in the image domain and then predict the latent label by directly applying the estimated representation coefficients in the label domain. However, due to the large gap between these two domains, the estimated representation coefficients in the image domain may not stay optimal for the label fusion. To overcome this dilemma, we propose a novel label fusion framework to make the weighting coefficients eventually to be optimal for the label fusion by progressively constructing a dynamic dictionary in a layer-by-layer manner, where a sequence of intermediate patch dictionaries gradually encode the transition from the patch representation coefficients in image domain to the optimal weights for label fusion. Our proposed framework is general to augment the label fusion performance of the current state-of-the-art methods. In our experiments, we apply our proposed method to hippocampus segmentation on ADNI dataset and achieve more accurate labeling results, compared to the counterpart methods with single-layer dictionary.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages190-197
Number of pages8
Volume9351
ISBN (Print)9783319245737
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015 - Munich, Germany
Duration: 2015 Oct 52015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9351
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015
CountryGermany
CityMunich
Period15/10/515/10/9

Fingerprint

Atlas
Glossaries
Labels
Fusion
Fusion reactions
Segmentation
Patch
Coefficient
Hippocampus
Neuroscience
Dilemma
Medical Image
Labeling
Weighting
Framework
Dictionary
Predict
Target
Experiment

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Song, Y., Wu, G., Sun, Q., Bahrami, K., Li, C., & Shen, D. (2015). Progressive label fusion framework for multi-atlas segmentation by dictionary evolution. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9351, pp. 190-197). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351). Springer Verlag. https://doi.org/10.1007/978-3-319-24574-4_23

Progressive label fusion framework for multi-atlas segmentation by dictionary evolution. / Song, Yantao; Wu, Guorong; Sun, Quansen; Bahrami, Khosro; Li, Chunming; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. p. 190-197 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Song, Y, Wu, G, Sun, Q, Bahrami, K, Li, C & Shen, D 2015, Progressive label fusion framework for multi-atlas segmentation by dictionary evolution. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9351, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9351, Springer Verlag, pp. 190-197, 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015, Munich, Germany, 15/10/5. https://doi.org/10.1007/978-3-319-24574-4_23
Song Y, Wu G, Sun Q, Bahrami K, Li C, Shen D. Progressive label fusion framework for multi-atlas segmentation by dictionary evolution. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351. Springer Verlag. 2015. p. 190-197. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-24574-4_23
Song, Yantao ; Wu, Guorong ; Sun, Quansen ; Bahrami, Khosro ; Li, Chunming ; Shen, Dinggang. / Progressive label fusion framework for multi-atlas segmentation by dictionary evolution. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. pp. 190-197 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{7b2b7476308c4fa5a9b706e6551f888a,
title = "Progressive label fusion framework for multi-atlas segmentation by dictionary evolution",
abstract = "Accurate segmentation of anatomical structures in medical images is very important in neuroscience studies. Recently, multi-atlas patch-based label fusion methods have achieved many successes, which generally represent each target patch from an atlas patch dictionary in the image domain and then predict the latent label by directly applying the estimated representation coefficients in the label domain. However, due to the large gap between these two domains, the estimated representation coefficients in the image domain may not stay optimal for the label fusion. To overcome this dilemma, we propose a novel label fusion framework to make the weighting coefficients eventually to be optimal for the label fusion by progressively constructing a dynamic dictionary in a layer-by-layer manner, where a sequence of intermediate patch dictionaries gradually encode the transition from the patch representation coefficients in image domain to the optimal weights for label fusion. Our proposed framework is general to augment the label fusion performance of the current state-of-the-art methods. In our experiments, we apply our proposed method to hippocampus segmentation on ADNI dataset and achieve more accurate labeling results, compared to the counterpart methods with single-layer dictionary.",
author = "Yantao Song and Guorong Wu and Quansen Sun and Khosro Bahrami and Chunming Li and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-24574-4_23",
language = "English",
isbn = "9783319245737",
volume = "9351",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "190--197",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Progressive label fusion framework for multi-atlas segmentation by dictionary evolution

AU - Song, Yantao

AU - Wu, Guorong

AU - Sun, Quansen

AU - Bahrami, Khosro

AU - Li, Chunming

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - Accurate segmentation of anatomical structures in medical images is very important in neuroscience studies. Recently, multi-atlas patch-based label fusion methods have achieved many successes, which generally represent each target patch from an atlas patch dictionary in the image domain and then predict the latent label by directly applying the estimated representation coefficients in the label domain. However, due to the large gap between these two domains, the estimated representation coefficients in the image domain may not stay optimal for the label fusion. To overcome this dilemma, we propose a novel label fusion framework to make the weighting coefficients eventually to be optimal for the label fusion by progressively constructing a dynamic dictionary in a layer-by-layer manner, where a sequence of intermediate patch dictionaries gradually encode the transition from the patch representation coefficients in image domain to the optimal weights for label fusion. Our proposed framework is general to augment the label fusion performance of the current state-of-the-art methods. In our experiments, we apply our proposed method to hippocampus segmentation on ADNI dataset and achieve more accurate labeling results, compared to the counterpart methods with single-layer dictionary.

AB - Accurate segmentation of anatomical structures in medical images is very important in neuroscience studies. Recently, multi-atlas patch-based label fusion methods have achieved many successes, which generally represent each target patch from an atlas patch dictionary in the image domain and then predict the latent label by directly applying the estimated representation coefficients in the label domain. However, due to the large gap between these two domains, the estimated representation coefficients in the image domain may not stay optimal for the label fusion. To overcome this dilemma, we propose a novel label fusion framework to make the weighting coefficients eventually to be optimal for the label fusion by progressively constructing a dynamic dictionary in a layer-by-layer manner, where a sequence of intermediate patch dictionaries gradually encode the transition from the patch representation coefficients in image domain to the optimal weights for label fusion. Our proposed framework is general to augment the label fusion performance of the current state-of-the-art methods. In our experiments, we apply our proposed method to hippocampus segmentation on ADNI dataset and achieve more accurate labeling results, compared to the counterpart methods with single-layer dictionary.

UR - http://www.scopus.com/inward/record.url?scp=84951754976&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951754976&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-24574-4_23

DO - 10.1007/978-3-319-24574-4_23

M3 - Conference contribution

SN - 9783319245737

VL - 9351

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 190

EP - 197

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -